DeepJ: Graph Convolutional Transformers with Differentiable Pooling for Patient Trajectory Modeling

2025-06-18

AI Summary

Existing graph-based methods for modeling electronic health records (EHRs) struggle to jointly capture intra-visit event interactions and inter-visit temporal dependencies, exhibiting static limitations—particularly in identifying clinically meaningful longitudinal event clusters. To address this, we propose the first dynamic graph modeling framework integrating Graph Convolutional Networks (GCNs) with a position-aware Transformer, augmented by a differentiable graph pooling mechanism that automatically discovers temporally and functionally coherent inter-visit event subgraphs. Our approach constructs an EHR temporal graph, enabling joint structural and sequential modeling of patients’ longitudinal clinical trajectories. In risk stratification tasks, it significantly outperforms five state-of-the-art baseline models while enhancing prediction interpretability: it precisely identifies prognostic-driving event subgraphs, offering a novel paradigm for clinical decision support.

Technology Category

• Machine Learning: Graph-based Machine Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Reasoning under Uncertainty: Graphical Models

Application Category

• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization

Abstract

In recent years, graph learning has gained significant interest for modeling complex interactions among medical events in structured Electronic Health Record (EHR) data. However, existing graph-based approaches often work in a static manner, either restricting interactions within individual encounters or collapsing all historical encounters into a single snapshot. As a result, when it is necessary to identify meaningful groups of medical events spanning longitudinal encounters, existing methods are inadequate in modeling interactions cross encounters while accounting for temporal dependencies. To address this limitation, we introduce Deep Patient Journey (DeepJ), a novel graph convolutional transformer model with differentiable graph pooling to effectively capture intra-encounter and inter-encounter medical event interactions. DeepJ can identify groups of temporally and functionally related medical events, offering valuable insights into key event clusters pertinent to patient outcome prediction. DeepJ significantly outperformed five state-of-the-art baseline models while enhancing interpretability, demonstrating its potential for improved patient risk stratification.

Problem

• Modeling dynamic medical event interactions across longitudinal EHR encounters
• Identifying temporally and functionally related medical event groups
• Improving patient risk stratification with interpretable event clusters

Innovation

• Graph convolutional transformers for EHR modeling
• Differentiable pooling captures cross-encounter interactions
• Identifies temporally related medical event clusters